Clinical neuroscientists have recently become interested in laboratory based simulated gambling tasks for investigating decision making deficits in drug abusers. Several studies have now shown that drug abusers perform poorly in these tasks compared to non abusers. These tasks have the potential to reveal interrelationships of several important factors, such as learning, responsivity to rewards and punishments, personality, and drug abuse profiles, which must be understood to enhance the treatment and prevention of drug abuse. Although useful for examining decision making in the laboratory, the behavior manifested in these tasks is multiply determined, making it impossible to isolate the component processes by simply quantifying the observable behaviors. For example, it is difficult to infer whether poor performance is related to insufficient attention to punishments, altered effects of rewards, inability to learn outcome contingencies, or simply reckless, random behavior in drug abusers. This is a translational research project in which mathematical models previously developed and tested in cognitive psychology will be applied to laboratory gambling tasks currently being used to study decision making in drug abusers. For the proposed project, formal models will be developed to reveal the critical processes that are intertwined and conflated within two gambling tasks. These models will provide a set of theoretically derived measures that describe and quantify basic cognitive processes that underlie the empirical data. We will use these new measures to determine the relative roles of learning, rewards, and punishments in the decision making behavior of drug abusers. Finally, we will determine the relationships between individual differences in drug abusers (i.e., type of drug abused, severity of current abuse, risk-seeking behavior traits) and the cognitive processes that underlie their decision making. The final product of this research will be an integration of our findings on decision making in drug abusers with available neuropsychological models of drug addiction. These theoretical analyses will permit a more direct evaluation of current neuropsychological theories by providing direct estimates of the mediating psychological constructs, thereby eliminating the need to making inferences from observable behavior to neurophysiology. Instead, it will finally be possible to specifically determine the implications of hypothesized neural systems for cognitive and motivational processes involved in drug abuse.